We form an academic-industrial collaboration to investigate and create a clinically translatable solution that accounts for a poorly-understood aspect of pulmonary toxicity in lung stereotactic ablative radiotherapy (SAbR) - radiation injury to branching serial structures (BSS), i.e., airways and pulmonary vessels. Lung SAbR involves the precise administration of very high, biologically potent doses (54-70 Gy) in relatively few fractions. While this highly successful technique has been demonstrated to achieve excellent 5-year local control (>90%), the use of such potent doses puts patients at risk for collateral toxiciy including radiation pneumonitis and radiation injury to airways, causing stenosis, atelectasis and ultimately fibrosis. A common limitation of current treatment planning strategies is that they use a relatively crude model of the lung as a uniform, solely parallel organ. There is compelling clinical evidence to show that this simplistic approach has limited power to predict and/or avoid toxicity. In order to address this gap in current knowledge, we propose the integration of virtual bronchoscopy technology into the radiation treatment planning process to map BSS segments, quantify their radiosensitivity and create treatment plans that limit dose to these structures. We hypothesize that anatomically variable radiation injury to the elements of the bronchial tree and pulmonary vasculature is an important determinant of post-SAbR toxicity and residual pulmonary function. We test this hypothesis through Aims 1-3 and develop a clinical translation framework for end-user evaluation of a prototype system in Aim 4. In Aim 1, we will perform a prospective clinical study with 40 lung cancer patients to assess the relationship between dose and radiation injury to BSS segments. Broncus will adapt their virtual bronchoscopy algorithms to create LungPointRT. This software will enable BSS autosegmentation and DICOMRT export. We will compute the dose to each segment and compare pre-SAbR and 8-12 months post-SAbR CT scans to assess radiation-induced segmental collapse. In Aim 2, we will acquire pre- and post-SAbR ventilation/perfusion (V/Q) SPECT-CT scans to spatially map the localized loss of pulmonary function, and determine the association between segmental collapse (Aim 1) and localized functional loss (Aim 2). These data will be used to estimate dose thresholds for each segment type. In Aim 3, we will develop novel treatment planning strategies that account for BSS radiosensitivity. In order to account for the increased complexity of the optimization problem, we will investigate parallelized global optimization implemented on graphic processor unit (GPU) platform. The optimization algorithms will be integrated into a research version of a clinical treatment planning system (TPS), Eclipse. In Aim 4, we will create a pre-clinical prototype system for end-user evaluation. The TPS and the LungpointRT will be installed on a GPU workstation. We will form an end-user evaluation team consisting of a physician, physicist and dosimetrist. The team will work with developers to iteratively refine user interfaces and clinical workflow, and to develop practice guidelines and education frameworks to facilitate clinical translation.